Systematic review and evaluation of methods of assessing urinary incontinence

OBJECTIVES: To identify and synthesise studies of diagnostic processes of urinary incontinence and to construct an economic model to examine the cost-effectiveness of simple, commonly used primary care tests. DATA SOURCES: The electronic databases MEDLINE (1966--2002), CINAHL (1982--2002) and EMBASE (1980--2002). REVIEW METHODS: Studies were selected and assessed using the Quality Assessment of Diagnostic Studies (QUADAS) tool. Studies that reported the results of applying the same diagnostic procedure using the same threshold value (cut-off) were pooled using a random effects meta-analysis model to produce pooled estimates of sensitivity, specificity and diagnostic odds ratio together with 95% confidence intervals. RESULTS: In total, 6009 papers were identified from the literature search, of which 129 were deemed relevant for inclusion in the review, and these papers compared two or more diagnostic techniques. The gold-standard diagnostic test for urinary incontinence with which each reference test was compared was multichannel urodynamics. In general, reporting in the primary studies was poor; there was a lack of literature in the key clinical areas and minimal literature dealing with diagnosis in men. Only a limited number of studies could be combined or synthesised, providing the following results when compared with multichannel urodynamics. A clinical history for diagnosing urodynamic stress incontinence (USI) in women was found to have a sensitivity of 0.92 and specificity of 0.56 and for detrusor overactivity (DO) a sensitivity of 0.61 and specificity of 0.87. For validated scales, question 3 of the Urogenital Distress Inventory was found to have a sensitivity of 0.88 and specificity of 0.60. Seven studies compared a pad test with multichannel urodynamics; however, four different pad tests were studied and therefore it was difficult to draw any conclusions about diagnostic accuracy. Of the four studies comparing urinary diary with multichannel urodynamics, only one presented data in a format that allowed sensitivity and specificity to be calculated. Their reported values of 0.88 and 0.83 suggest that a urinary diary may be effective in the diagnosis of DO in women. Examination of the incremental cost-effectiveness of three primary care tests used in addition to history found that the diary had the lowest cost-effectiveness ratio of between pound 35 and pound 77 per extra unit of effectiveness (or case diagnosed). Imaging by ultrasound to determine leakage was found to be effective in the diagnosis of USI in women, with a sensitivity of 0.94 and specificity of 0.83. CONCLUSIONS: This is the first systematic review of methods for diagnosing urinary incontinence. As reporting of the primary studies was poor, clinical interpretation was often difficult because few studies could be synthesised and conclusions made. The report found that a large proportion of women with USI can be correctly diagnosed in primary care from clinical history alone. On the basis of diagnosis the diary appears to be the most cost-effective of the three primary care tests (diary, pad test and validated scales) used in addition to clinical history. Ultrasound imaging may offer a valuable alternative to urodynamic investigation. The clinical stress test is effective in the diagnosis of USI. Adaptation of such a test so that it could be performed in primary care with a naturally filled bladder may prove clinically useful. If a patient is to undergo an invasive urodynamic procedure, multichannel urodynamics is likely to give the most accurate result in a secondary care setting. There is a dearth of literature on the diagnosis of urinary incontinence in men, with no studies meeting the study criteria for data extraction in the diagnosis of bladder outlet obstruction. There is a need for large-scale, high-quality primary studies evaluating the use of a number of diagnostic methods in a primary care setting to be undertaken so that the results of this systematic review can be verified or not. Such studies should include not only an assessment of clinical effectiveness, in this case diagnostic accuracy, but also an assessment of costs and quality of life/satisfaction to inform future health policy decisions. Studies carried out should be reported to a better standard. The recommendations of the Standards for Reporting Diagnostic Accuracy (STARD) initiative should be followed to ensure the accuracy and completeness of reporting design and results.